Interstrand Aminoacyl Transfer in a tRNA Acceptor Stem-Overhang Mimic

Abstract

Protein-catalyzed aminoacylation of the 3'-overhang of tRNA by an aminoacyl-adenylate could not have taken place prior to the advent of genetically coded peptide synthesis, and yet the latter process has an absolute requirement for aminoacyl-tRNA. There must therefore have been an earlier nonprotein-catalyzed means of generating aminoacyl-tRNA. Here, we demonstrate efficient interstrand aminoacyl transfer from an aminoacyl phosphate mixed anhydride at the 5'-terminus of a tRNA acceptor stem mimic to the 2',3'-diol terminus of a short 3'-overhang. With certain five-base 3'-overhangs, the transfer of an alanyl residue is highly stereoselective with the l-enantiomer being favored to the extent of ∼10:1 over the d-enantiomer and is much more efficient than the transfer of a glycyl residue. N-Acyl-aminoacyl residues are similarly transferred from a mixed anhydride with the 5'-phosphate to the 2',3'-diol but with a different dependence of efficiency and stereoselectivity on the 3'-overhang length and sequence. Given a prebiotically plausible and compatible synthesis of aminoacyl phosphate mixed anhydrides, these results suggest that RNA molecules with acceptor stem termini resembling modern tRNAs could have been spontaneously aminoacylated, in a stereoselective and chemoselective manner, at their 2',3'-diol termini prior to the onset of protein-catalyzed aminoacylation.

Figure 1

Model for the origin of tRNA by duplication suggests the possibility of interstrand aminoacyl transfer at the acceptor stem overhang. (a) Cloverleaf 2D structure of extant tRNA with the anticodon loop and extended conformation 3′-overhang highlighted (dark blue). (b) Two copies of RNA annealed to each other to generate a cloverleaf proto-tRNA with two folded-back conformation 3′-overhangs (dark blue). (c) Close up of the stem 3′-overhang at the bottom of the structure depicted in b showing the proximity that would be required for attack of the 3′-hydroxyl group on the activated 5′-phosphate (black) to generate a loop. (d) Close up of the stem 3′-overhang at the top of the structure depicted in b showing how the same proximity shown in c might allow transfer of an aminoacyl group from a mixed anhydride with the 5′-phosphate (black) to the 3′-hydroxyl group of the folded-back conformation 3′-overhang.

Figure 2

Characterization of nicked loop l-alanyl-transfer. (a) MALDI-TOF mass spectrum of products of aminoacyl transfer from 5′-l-Ala-pAGCGA-3′ to an acceptor strand with sequence 5′-UCGCUUGCCA-3′. Mass clusters for the acceptor strand (found 3096.885, calcd 3096.44) and its l-alanyl diol ester product (5′-UCGCUUGCCA-l-Ala, found 3167.935, calcd 3167.44) are indicated. Measured mass difference, 71.050, C₃H₅NO [Ala-H₂O], calcd 71.037. (b) Enzyme digestion confirming that the diol of the acceptor strand’s 3′-terminal adenosine is aminoacylated by aminoacyl transfer. (c) l-Ala transfer in a tRNA acceptor arm mimic: 5′-UCGCUUUCCA-3′; 5′-l-Ala-pAGCGA-3′. Peaks for the donor, donor mixed anhydride, and acceptor strands are indicated. Peak presumed to be due to the diol ester transfer product is highlighted by the dashed box. (d) Time courses showing the stereoselectivity for l-Ala over d-Ala nicked loop transfer with acceptor strands UCGCUUGCCA and UCGCUUUCCA. Conditions: each oligoribonucleotide 100 μM, NaCl 100 mM, MgCl₂ 5 mM, HEPES 50 mM, pH 6.8, 10 °C.